- 05 4月, 2013 16 次提交
-
-
由 Peter Krempa 提交于
The last Viktor's effort to fix the race and memory corruption unfortunately wasn't complete in the case the close callback was not registered in an connection. At that time, the trail of event's that I'll describe later could still happen and corrupt the memory or cause a crash of the client (including the daemon in case of a p2p migration). Consider the following prerequisities and trail of events: Let's have a remote connection to a hypervisor that doesn't have a close callback registered and the client is using the event loop. The crash happens in cooperation of 2 threads. Thread E is the event loop and thread W is the worker that does some stuff. R denotes the remote client. 1.) W - The client finishes everything and sheds the last reference on the client 2.) W - The virObject stuff invokes virConnectDispose that invokes doRemoteClose 3.) W - the remote close method invokes the REMOTE_PROC_CLOSE RPC method. 4.) W - The thread is preempted at this point. 5.) R - The remote side receives the close and closes the socket. 6.) E - poll() wakes up due to the closed socket and invokes the close callback 7.) E - The event loop is preempted right before remoteClientCloseFunc is called 8.) W - The worker now finishes, and frees the conn object. 9.) E - The remoteClientCloseFunc accesses the now-freed conn object in the attempt to retrieve pointer for the real close callback. 10.) Kaboom, corrupted memory/segfault. This patch tries to fix this by introducing a new object that survives the freeing of the connection object. We can't increase the reference count on the connection object itself or the connection would never be closed, as the connection is closed only when the reference count reaches zero. The new object - virConnectCloseCallbackData - is a lockable object that keeps the pointers to the real user registered callback and ensures that the connection callback is either not called if the connection was already freed or that the connection isn't freed while this is being called.
-
由 Peter Krempa 提交于
This patch improves the error message after disconnecting from the hypervisor and adds the close callback operations required not to leak the callback reference.
-
由 Peter Krempa 提交于
The function is used to establish connection so it should be in the main virsh file. This movement also enables further improvements done in next patches. Note that the "connect" command has moved from the host section of virsh to the main section. It is now listed by 'virsh help virsh' instead of 'virsh help host'.
-
由 Viktor Mihajlovski 提交于
Before closing the connection we unregister the close callback to prevent a reference leak. Further, the messages on virConnectClose != 0 are a bit more specific now. Signed-off-by: NViktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
-
由 Viktor Mihajlovski 提交于
By adjusting the reference count of the connection object we prevent races between callback function and virConnectClose. Signed-off-by: NViktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
-
由 Peter Krempa 提交于
This patch refactors various places to allow removing of the defaultConsoleTargetType callback from the virCaps structure. A new console character device target type is introduced - VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_NONE - to mark that no type was specified in the XML. This type is at the end converted to the standard VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL. Other types that are different from this default have to be processed separately in the device post parse callback.
-
由 Peter Krempa 提交于
Use the virDomainXMLConf structure to hold this data and tweak the code to avoid semantic change. Without configuration the KVM mac prefix is used by default. I chose it as it's in the privately administered segment so it should be usable for any purposes.
-
由 Peter Krempa 提交于
Use the virDomainXMLConf structure to hold this data.
-
由 Peter Krempa 提交于
Use the qemu specific callback to fill this data in the qemu driver as it's the only place where it was used and fix tests as the qemu test capability object didn't configure the defaults for the tests.
-
由 Peter Krempa 提交于
This patch removes the emulatorRequired field and associated infrastructure from the virCaps object. Instead the driver specific callbacks are used as this field isn't enforced by all drivers. This patch implements the appropriate callbacks in the qemu and lxc driver and moves to check to that location.
-
由 Peter Krempa 提交于
This patch removes the defaultDiskDriverName from the virCaps structure. This particular default value is used only in the qemu driver so this patch uses the recently added callback to fill the driver name if it's needed instead of propagating it through virCaps.
-
由 Peter Krempa 提交于
This gets rid of the parameter in favor of using the new callback infrastructure to do the same stuff. This patch implements the domain adjustment callback in the openVZ driver and moves the check from the parser to a new validation method in the callback infrastructure.
-
由 Peter Krempa 提交于
This patch implements the devices post parse callback and uses it to fill the default qemu network card model into the XML if none is specified. Libvirt assumes that the network card model for qemu is the "rtl8139". Record this in the XML using the new callback to avoid user confusion.
-
由 Peter Krempa 提交于
Move the xmlopt and caps arguments to the end of the argument list.
-
由 Peter Krempa 提交于
This patch adds instrumentation that will allow hypervisor drivers to fill and validate domain and device definitions after parsed by the XML parser. With this patch, after the XML is parsed, a callback to the driver is issued requesting to fill and validate driver specific details of the configuration. This allows to use sensible defaults and checks on a per driver basis at the time the XML is parsed. Two callback pointers are stored in the new virDomainXMLConf object: * virDomainDeviceDefPostParseCallback (devicesPostParseCallback) - called for a single device parsed and for every single device in a domain config. A virDomainDeviceDefPtr is passed along with the domain definition and virCaps. * virDomainDefPostParseCallback, (domainPostParseCallback) - A callback that is meant to process the domain config after it's parsed. A virDomainDefPtr is passed along with virCaps. Both types of callbacks support arbitrary opaque data passed for the callback functions. Errors may be reported in those callbacks resulting in a XML parsing failure.
-
由 Peter Krempa 提交于
This patch is the result of running: for i in $(git ls-files | grep -v html | grep -v \.po$ ); do sed -i -e "s/virDomainXMLConf/virDomainXMLOption/g" -e "s/xmlconf/xmlopt/g" $i done and a few manual tweaks.
-
- 04 4月, 2013 13 次提交
-
-
由 Daniel P. Berrange 提交于
If libnuma is not compiled in, or numa_available() returns an error, stub out fake NUMA info consisting of one NUMA cell containing all CPUs and memory. Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
Not all kernel builds have any entries under the location /sys/devices/system/cpu/cpu0/topology. We already cope with that being missing in some cases, but not all. Update the code which looks for thread_siblings to cope with the missing file Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
The Raspberry Pi runs the armv6l architecture and apparently people are trying to run libvirt LXC on it. So we should allow that as a valid arch Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
Implement the bare minimal sysinfo for ARM platforms by reading the CPU models from /proc/cpuinfo Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
Certain functions in the sysinfotest.c are not used unless a whitelisted architecture is being built. Disable those functions unless required to avoid warnings about unused functions. sysinfotest.c:93:1: warning: 'sysinfotest_run' defined but not used [-Wunused-function] sysinfotest_run(const char *test, Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
The sysinfo code used by QEMU is trivially portable to the LXC driver
-
由 Daniel P. Berrange 提交于
The nodedev driver currently only detects harddisk, cdrom and floppy devices. This adds support for SD cards, which are common storage for ARM devices, eg the Google ChromeBook <device> <name>block_mmcblk0_0xb1c7c08b</name> <parent>computer</parent> <capability type='storage'> <block>/dev/mmcblk0</block> <drive_type>sd</drive_type> <serial>0xb1c7c08b</serial> <size>15758000128</size> <logical_block_size>512</logical_block_size> <num_blocks>30777344</num_blocks> </capability> </device> Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
This reverts commit c9c87376. Now that we force all containers to have a root filesystem, there is no way the host's /dev is ever exposed Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
Currently the LXC container code has two codepaths, depending on whether there is a <filesystem> element with a target path of '/'. If we automatically add a <filesystem> device with src=/ and dst=/, for any container which has not specified a root filesystem, then we only need one codepath for setting up the filesystem. Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
Early on kernel support for private devpts was not widespread, so we had compatibiltiy codepaths. Such old kernels are not seriously used for LXC these days, so the compat code can go away Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Atsushi Kumagai 提交于
When creating a logical volume with virStorageVolCreateXMLFrom, "qemu-img convert" is called internally if clonevol is a file volume. Then, vol->target.format is used as output_fmt parameter but the target.format of logical volumes is always 0 because logical volumes haven't the volume format type element. Fortunately, 0 was treated as RAW file format before commit f772b3d9, so there was no problem. But now, 0 is treated as the type of none, qemu-img fails with "Unknown file format 'none'". This patch fixes this issue by treating output block devices as RAW file format like for input block devices. Signed-off-by: NAtsushi Kumagai <kumagai-atsushi@mxc.nes.nec.co.jp>
-
由 Guido Günther 提交于
otherwise we crash later on if we don't find a match like: #0 0xb72c2b4f in virSecurityManagerGenLabel (mgr=0xb8e42d20, vm=0xb8ef40c0) at security/security_manager.c:424 #1 0xb18811f3 in qemuProcessStart (conn=conn@entry=0xb8eed880, driver=driver@entry=0xb8e3b1e0, vm=vm@entry=0xb8ef58f0, migrateFrom=migrateFrom@entry=0xb18f6088 "stdio", stdin_fd=18, stdin_path=stdin_path@entry=0xb8ea7798 "/var/lib/jenkins/jobs/libvirt-tck-build/workspace/tck.img", snapshot=snapshot@entry=0x0, vmop=vmop@entry=VIR_NETDEV_VPORT_PROFILE_OP_RESTORE, flags=flags@entry=2) at qemu/qemu_process.c:3364 #2 0xb18d6cb2 in qemuDomainSaveImageStartVM (conn=conn@entry=0xb8eed880, driver=driver@entry=0xb8e3b1e0, vm=0xb8ef58f0, fd=fd@entry=0xb6bf3f98, header=header@entry=0xb6bf3fa0, path=path@entry=0xb8ea7798 "/var/lib/jenkins/jobs/libvirt-tck-build/workspace/tck.img", start_paused=start_paused@entry=false) at qemu/qemu_driver.c:4843 #3 0xb18d7eeb in qemuDomainRestoreFlags (conn=conn@entry=0xb8eed880, path=path@entry=0xb8ea7798 "/var/lib/jenkins/jobs/libvirt-tck-build/workspace/tck.img", dxml=dxml@entry=0x0, flags=flags@entry=0) at qemu/qemu_driver.c:4962 #4 0xb18d8123 in qemuDomainRestore (conn=0xb8eed880, path=0xb8ea7798 "/var/lib/jenkins/jobs/libvirt-tck-build/workspace/tck.img") at qemu/qemu_driver.c:4987 #5 0xb718d186 in virDomainRestore (conn=0xb8eed880, from=0xb8ea87d8 "/var/lib/jenkins/jobs/libvirt-tck-build/workspace/tck.img") at libvirt.c:2768 #6 0xb7736363 in remoteDispatchDomainRestore (args=<optimized out>, rerr=0xb6bf41f0, client=0xb8eedaf0, server=<optimized out>, msg=<optimized out>) at remote_dispatch.h:4679 #7 remoteDispatchDomainRestoreHelper (server=0xb8e1a3e0, client=0xb8eedaf0, msg=0xb8ee72c8, rerr=0xb6bf41f0, args=0xb8ea8968, ret=0xb8ef5330) at remote_dispatch.h:4661 #8 0xb720db01 in virNetServerProgramDispatchCall (msg=0xb8ee72c8, client=0xb8eedaf0, server=0xb8e1a3e0, prog=0xb8e216b0) at rpc/virnetserverprogram.c:439 #9 virNetServerProgramDispatch (prog=0xb8e216b0, server=server@entry=0xb8e1a3e0, client=0xb8eedaf0, msg=0xb8ee72c8) at rpc/virnetserverprogram.c:305 #10 0xb7206e97 in virNetServerProcessMsg (msg=<optimized out>, prog=<optimized out>, client=<optimized out>, srv=0xb8e1a3e0) at rpc/virnetserver.c:162 #11 virNetServerHandleJob (jobOpaque=0xb8ea7720, opaque=0xb8e1a3e0) at rpc/virnetserver.c:183 #12 0xb70f9f78 in virThreadPoolWorker (opaque=opaque@entry=0xb8e1a540) at util/virthreadpool.c:144 #13 0xb70f94a5 in virThreadHelper (data=0xb8e0e558) at util/virthreadpthread.c:161 #14 0xb705d954 in start_thread (arg=0xb6bf4b70) at pthread_create.c:304 #15 0xb6fd595e in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:130 This unbreaks libvirt-tck's domain/100-transient-save-restore.t with qemu:///session and selinux compiled in but disabled. Introduced by 8d68cbea
-
由 Eric Blake 提交于
Commit f84b92ea introduced a memory leak on error; John Ferlan reported that valgrind caught it during 'make check'. * src/qemu/qemu_command.c (qemuBuildMachineArgStr): Plug leak.
-
- 03 4月, 2013 9 次提交
-
-
由 Daniel P. Berrange 提交于
By passing the flags -z relro -z now to the linker, we can force it to resolve all library symbols at startup, instead of on-demand. This allows it to then make the global offset table (GOT) read-only, which makes some security attacks harder. Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Daniel P. Berrange 提交于
PIE (position independent executable) adds security to executables by composing them entirely of position-independent code (PIC. The .so libraries already build with -fPIC. This adds -fPIE which is the equivalent to -fPIC, but for executables. This for allows Exec Shield to use address space layout randomization to prevent attackers from knowing where existing executable code is during a security attack using exploits that rely on knowing the offset of the executable code in the binary, such as return-to-libc attacks. Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
-
由 Peter Krempa 提交于
The JSON generator is able to represent only values less than LLONG_MAX, fix the bandwidth limit checks when converting to value to catch overflows before they reach the generator.
-
由 Osier Yang 提交于
It's long enough to have a independant definition.
-
由 Daniel P. Berrange 提交于
Every source file is currently built twice by libtool, once for the shared library and once for the static library. Static libs are not commonly packaged by distros and slow down compilation time by more than 50% compared to a shared-only build time. Time for 'make -j 4': shared only: 2 mins 9 secs shared + static: 3 mins 26 secs Time for non-parallel make shared only: 3 mins 32 secs shared + static: 5 mins 41 secs Those few people who really want them, can pass --enable-static to configure Disabling them by default requires use of LT_INIT, but for compat with RHEL5 we can't rely on that. So we conditionally use LT_INIT, but fallback to AM_PROG_LIBTOOL if not present.
-
由 Peter Krempa 提交于
s/persitent/persistent/
-
由 Michal Privoznik 提交于
https://bugzilla.redhat.com/show_bug.cgi?id=947387 If a user configures a domain to use a seclabel of a specific type, but the appropriate driver is not accessible, we should refuse to start the domain. For instance, if user requires selinux, but it is either non present in the system, or is just disabled, we should not start the domain. Moreover, since we are touching only those labels we have a security driver for, the other labels may confuse libvirt when reconnecting to a domain on libvirtd restart. In our selinux example, when starting up a domain, missing security label is okay, as we auto-generate one. But later, when libvirt is re-connecting to a live qemu instance, we parse a state XML, where security label is required and it is an error if missing: error : virSecurityLabelDefParseXML:3228 : XML error: security label is missing This results in a qemu process left behind without any libvirt control.
-
由 Martin Kletzander 提交于
virsh schedinfo was able to set only one parameter at a time (not counting the deprecated options), but it is useful to set more at once, so this patch adds the possibility to do stuff like this: virsh schedinfo <domain> cpu_shares=0 vcpu_period=0 vcpu_quota=0 \ emulator_period=0 emulator_quota=0 Invalid scheduler options are reported as well. These were previously reported only if the command hadn't updated any values (when cmdSchedInfoUpdate returned 0). Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=810078 Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=919372 Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=919375
-
由 Peter Krempa 提交于
Mimic the fix done in 02b90972 to fix crash by accessing an already freed structure. Also copy the explaining comment why the pointer can't be accessed any more.
-
- 02 4月, 2013 2 次提交
-
-
由 Michal Privoznik 提交于
https://bugzilla.redhat.com/show_bug.cgi?id=928197 The virsh domfstrim command was not freeing allocated domain, leaving leaked references behind.
-
由 Martin Kletzander 提交于
Descriptions for vol-download and vol-upload didn't make much sense. Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=923613
-